Contributors - Abhishek Pratap Singh, Kishore M
- This file train a single token based biased model.
- The weights after training are saved.
- This file is used for two task:
- To generate the probability distribution predicted by biased model for each token. This will be used for debiasing
- To generate predictions in terms of BIO tags for each token. This will be used for categorising the performance over various catergories of mentions